Cloud based linked data platform for Structural Engineering

advertisement
Cloud based linked data platform for
Structural Engineering Experiment
Xiaohui Zhang
[email protected]
Outline



Motivation
The CLDP-SEE Platform
Conclusion and Future Work
2
Motivation

Structural Engineering




A discipline analyzing the force and deformation of
buildings by mechanical methods.
Experiment is one of the main means for domain
research.
Large amounts of experimental data is accumulated,
but be maintained by each experimental user
dispersedly.
Due to the complexity and heterogeneity of the
experimental data, the sharing and integrating with
the traditional methods is difficult.
3
Motivation

Linked Data



Linked Data is simply about using the Web to
publish structured data and create typed links
between data from different sources.
Based on semantic web, linked data uses RDF to
make typed statements that link arbitrary things in
the world.
Linked data provides a wonderful approach to
publish and consume data on the web and make the
web be a global data space which can be
understood both by computer and human.
4
Motivation

Linked Data for Structure Engineering



The data represented based on semantic can be
understood by machines, which is helpful for
the integration and processing of experimental
data.
The interlinking among data from different
sources is a effective measure for the
heterogeneity.
Linked data will make it easy for the sharing
and intelligent processing of experimental data.
5
Motivation

A huge challenge for domain researchers to
deploy and use Linked Data related tools to
make operations on the data:




Conversion of data format
Publication of experiment data
Integration of experiment data
Consuming of linked experiment data
6
Motivation



A centralized platform providing all the functions
needed by experiencing linked data in services is
necessary for domain researchers.
A linked data platform based on cloud for Structural
Engineering Experiment (CLDP-SEE) is proposed by this
paper.
The publishing, interlinking and consuming of
experiment data is an intact ecosystem of data sharing.
CLDP-SEE can


lower the threshold of sharing data with linked data
technology for domain users;
promote the growth of the linked data ecosystem and the
development of Structural Engineering discipline.
7
The CLDP-SEE Platform
The application scenario of CLDP-SEE
8
The CLDP-SEE Platform

The operations in application scenario:





Uploading and managing the RDF data, setting access
control policies of each datasets.
Uploading raw data in traditional formats, such as CSV,
Excel, Relational Database. And then converting these
raw data into RDF.
Querying datasets from the shared data space, private
data space according to the authority and even the
datasets from the Web, and then interlinking data
among these datasets to generate a Virtual Data Space.
Reasoning and querying the data in Virtual Data Space.
Publishing data with Linked Data Server.
9
The CLDP-SEE Platform

The Architecture of CLDP-SEE
10
The CLDP-SEE Platform

Portal Layer

Provides graphical web interface for users to
experience almost all the functions providing by
CLDP-SEE.
11
The CLDP-SEE Platform

Core Service Layer

Data Manage Service: is mainly used to help
users to manage their data.






Data Upload
Data Format Transform
Dataset Registry
Dataset Manage
Data Publish
Authority Manage
12
The CLDP-SEE Platform

Core Service Layer

Data Link Service




Provides the capabilities of data integration;
Coreference Interlink is responsible for getting the
request of users, and finding the coreference relations
between data from different datasets.
The coreference relation of RDF data refers to two
different URI pointing to the same entity.
Two methods of coreference interlinking:



Similarity computation: implemented according to SILk(Isele,
R.; Jentzsch, A. & Bizer, C. 2010)
Rules matching: Link Rule Manage service provides graphical
interface for the experts and users to define rules.
Links Update will update the links with the information
collected by Dataset Monitor service.
13
The CLDP-SEE Platform

Core Service Layer

Data Reason Service




The rule-based inference is mainly done by this service.
Users can select any datasets from Virtual Data Space,
Private Space or Shared Data Space according to the
authority.
Inference Rule Manage supports each user to define and
manage their private inference rules, and check the
consistency with default rules provided by domain
experts.
Default rules and user-defined rules can be applied in the
inference.
14
The CLDP-SEE Platform

Core Service Layer

Data Query Service


The basis of consuming linked experiment data.
Two kinds of query interfaces:




navigation query based on SEE ontology
query based on keywords
Support users self-defining the scope of query.
Query Engine is responsible for processing the
request from self-service portal, and executes
SPARQL query on the datasets selected by users.
15
The CLDP-SEE Platform

Supporting Service Layer


The services in this layer are mainly supporting the
functions of the services in Core Service Layer.
Data service mainly provides the underlying
functions of RDF data management and access.


Ontology Manage service, Dataset Access service ,
Dataset Storage service, Dataset Monitor.
Publish Service mainly supports the Data Publish in
Data Manage Service.


Linked Data Server
RDF File Server
16
The CLDP-SEE Platform

Supporting Service Layer

User Service:



Metadata Manage service: manages the information
of users and make user can update personal
materials.
Role Manage service: be provided for platform
administrator to manage the roles of users.
Social Network Manage service: manages the friend
relationships among users, and provides personal
space for each user.
17
The CLDP-SEE Platform

Data Storage Layer

SEE Ontology


RDF Datasets


stores the datasets in users’ Private Data Space, and
ensure the isolation between users.
Links of Data


stores the unified ontology schema and the data in
Shared Data Space.
stores the relation between the entities from different
datasets.
Rule Base


default rule bases
user defined rules
18
Conclusion and Future Work


CLDP-SEE provides almost all the services
needed by Structural Engineering domain
users to manage and share experiment data
based on linked data technology.
Future work:


Improving the performance of data linking and
inference.
More flexible access control policy and finegrained access control model.
19
Thank You!
20
Related Works

Publication of Linked Data


D2R Server (Prud’hommeaux & Seaborne.
2006) :publishing the content of relational
databases as RDF.
Pubby and Elda: providing Linked Data
interfaces for RDF data sources.
21
Related Works

Searching and Browsing of Linked Data

linked data browser: enables people to view
data from one dataset to another by following
RDF links.




Tabulator (Berners-Lee et al., 2006)
OpenLink Browser
(http://oat.openlinksw.com/rdfbrowser2/)
Marbles (http://marbles.sourceforge.net/)
linked data engine: provides service for people
querying the Web of Data.

Falcons, Sindice, Swoogle and SWSE
22
Related Works

Interlinking of Linked Data




SILK (Robert et al., 2010)
DSNotify(Haslhofer & Popitsch, 2009)
LinkedDataBR (Kelli et al., 2011): a platform
used by Brazil for linking open Brazilian
governmental data.
Talis: a platform for RDF data sharing via
weaving data with the Web to create a highly
available and adaptable environment.
(http://www.talis.com/platform/)
23
Related Works


CLDP-SEE provides services for the storage,
query, publishing and management of RDF data.
CLDP-SEE provides more perfect services with
cloud characteristics:




More flexible and personalized self-service model;
Query the datasets according to subject, and
ineterlink the data in the result datasets;
Elastical reasoning service on the user-defined
datasets;
A shared RDF repository with rich interlinks among
data.
24
Download